Epidemiological data show that men are at higher risk for coronary heart disease (CHD) that are women and suggestive data indicate that female reproductive hormones vary with CHD risk. Behavioral and biological risk factors, at least as they are currently measured, cannot account for the sex differences in risk for CHD. We propose that to the extent that (1) individual differences in cardiovascular and neuroendocrine responses to behavioral stress are a risk factor for CHD, and (2) differences in cardiovascular and neuroendocrine responses to stress exist between men and women and among women who vary in reproductive hormones, (3) the differential risk for CHD among those groups are in part the result of corresponding differences in physiological responses to stress. Furthermore, we propose that (4) sex and reproductive hormonal status differences in physiological responses to stress, if obtained, are determined by an interaction between the characteristics of the behavioral stressors and level of reproductive hormones. The first and third hypotheses are epidemiological in nature and not addressed here. The others are open to experimental investigation and are tested by three experiments. The first study will evaluate the cardiovascular, metabolic, and neuroendocrine responses to behavioral stressors in regularly cycling women according to menstrual phase. The second study will evaluate those responses in age-matched men, premenopausal women, and postmenopausal women (both natural and surgical menopause) in the laboratory. To test the extent to which the laboratory results generalize, the level of stress and blood pressure and heart rates will also be measured during two consecutive days in the same individuals. The final study will be the first experimental test of the Frankenhaeuser hypothesis that sex differences in catecholamine responses are due to differences in sex roles. This will involve an experimental manipulation whereby the same task will be presented as calling upon masculine, feminine, or sex-neutral characteristics.